A cross-sectional analysis of low-cost non-cancer registry recruitment sources for a hybrid type 2 trial evaluating a digital intervention for melanoma survivors

Introduction

A major challenge in intervention research for cancer survivors is that many programs shown to be effective in research settings are not subsequently tested under real-world conditions, which limits accessibility and scalability and prevents efficacious interventions from reaching populations that may benefit. Cancer registries have traditionally served as a valuable resource for research recruitment. However, their use is accompanied by several limitations and challenges (1-3). These databases are often affected by issues such as costs, data incompleteness, duplicate case reports, reporting delays, and misclassification of race and ethnicity, all of which can compromise data collection and interpretation (2). To enhance data quality and reliability, additional staff training and regular data audits are required to meet established standards for registry data (3).

Researchers have employed a range of more novel recruitment strategies to reach and enroll a broader swath of participants in cancer-related intervention trials. These sources include social media (e.g., paid advertisements and volunteer influencer engagement), electronic health record (EHR) messaging, and healthcare provider network collaborations. The participant demographics and enrollment rate of these approaches often varies depending on the characteristics of the target population.

Although social media has been increasingly utilized for recruitment in cancer survivor intervention trials, evidence regarding its value remains mixed. For instance, a 2019 UK-based study successfully recruited 200 participants with a prior cancer diagnosis using platforms such as Facebook, Twitter, and Reddit for a longitudinal intervention exploring predictors for patient-reported outcomes in cancer survivors (4). The study found that the success of different online methods varied over time and that online recruitment produced a more demographically and clinically representative sample than in-person approaches (4). Similarly, another study, focused on female cancer survivors’ fertility experiences and post-treatment decision making, compared social media to hospital-based modes of recruitment (5). They found that social media yielded a higher enrollment rate (37%; n=54/146) compared to hospital recruitment (7%; n=21/289), while also requiring fewer resources (5). However, not all studies have supported social media as a recruitment approach. A 2016 systematic review of 30 studies found that social media was the most effective recruitment strategy in only 12 studies, while 15 identified other methods (e.g., community-based organizations, friend referrals) as more effective, and 3 found social media to be equally effective as another recruitment method in terms of enrollment (6). Evidence for social media recruitment approaches remains limited and mixed, with differences attributed to varying populations of interest, engagement strategies, and specific platform characteristics. These findings suggest the need for further investigation into the value of social media sources.

In addition to using social media, collaborating with provider networks has been effective for research recruitment (7-9). Lamberti and colleagues evaluated recruitment and retention strategies used by 98 studies within the Tufts Center for the Study of Drug Development and found that 42 (43%) utilized physician referrals (10). Among those, 10 studies (28.5%) identified physician referrals as one of the top three most effective methods for creating study awareness (10). Additionally, electronic health chart screening was used in 41 studies (41.8%), with 17 (41.5%) of these studies ranking it among the top three most effective study awareness strategies (10). Although research on physician engagement is limited, findings suggest that both referrals and chart prescreening can be effective recruitment strategies for research. This may be due to patients’ trust in their healthcare providers and provider networks, as physicians remain among the most trusted sources of health information (11).

Comparing varying recruitment sources is essential, as they likely influence enrollment, data quality, and sample representativeness. This study evaluated the enrollment rates and demographic characteristics from three recruitment sources (i.e., social media, EHR messaging, and dermatology provider network outreach) within a hybrid type 2 dissemination-effectiveness randomized trial, mySmartSkin.

The mySmartSkin intervention is an efficacious digital program for melanoma survivors designed to promote sun safety and skin self-examinations (SSEs) to prevent and detect melanoma recurrence (12). The intervention includes two interactive multimedia modules focused on sun safety and SSE, along with an SSE body map and monthly reminders to encourage regular self-examinations. Participants can record concerning or changing lesions on the body map and upload digital photographs for review with a healthcare provider. To support engagement, sun protection items (e.g., sunscreen, hats) and SSE-related tools (e.g., ruler to measure spots, reminder stickers) were provided as incentives after completion of each module and the first two SSEs. The comparison condition consisted of static, non-interactive online educational materials covering the same topics. Participants in both groups received e-gift cards for completing online surveys through 18 months. We present this article in accordance with the STROBE reporting checklist (available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-25-61/rc).

Methods

Eligibility criteria

Cutaneous malignant melanoma survivors were recruited to participate in an online behavioral intervention using a variety of recruitment strategies to broaden the program’s reach and assess real-world dissemination conditions (12). Eligibility criteria to participate in the study were: (I) diagnosis of primary pathologic stage 0–III cutaneous malignant melanoma; (II) 3 months to 5 years post-surgery; (III) no current evidence of melanoma; (IV) not adherent to thorough SSE (i.e., did not check entire body at least once during the past 3 months); (V) between 18 and 89 years old; (VI) access to the internet; (VII) able to speak and read English; and (VIII) able to provide informed consent.

Recruitment sources

Three non-cancer registry recruitment sources were employed to recruit convenience samples: (I) social media (e.g., paid Facebook advertisements, unpaid posts shared by melanoma-focused influencers and organizations on Facebook and Instagram); (II) EHR messaging (e.g., EPIC message blasts via medical center records); and (III) outreach within a national dermatology practice network. Of all deployed recruitment methods, Facebook advertisements were the only paid strategy; all others were implemented at no cost other than study staff time.

Social media

We collaborated with a communication agency to develop targeted advertisements for Facebook. To reach the eligible population, we targeted cities in states that had higher rates of melanoma, including Texas, Florida, Minnesota, New Hampshire, Utah, Idaho, and Vermont. Since the melanoma survivor population has an average age of approximately 60 years, advertisement imagery was selected to represent individuals from diverse racial and ethnic backgrounds within this age range, shown doing activities commonly associated with the targeted geographic regions (e.g., boating and fishing for coastal cities). Additionally, a higher proportion of advertisements featured male-presenting individuals, based on prior recruitment experience indicating lower enrollment rates among men in melanoma-related research. The initial ads were deployed using Meta Ads, a platform that provides detailed performance metrics such as reach, impressions, and click-through rates. Initial “brand awareness” ads were deployed over a 7-week period, allowing sufficient time to gather meaningful engagement data. Based on this feedback, we refined key elements, including model demographics in ad images, language and messaging, to enhance relevance and appeal. With the revised ads, we deployed ad campaigns focused on “traffic” and engagement (Table S1).

We also reached out to a variety of local and national melanoma and skin cancer influencers (e.g., melanoma survivors, family members of patients) and organizations to voluntarily share our advertisements across their platforms (e.g., Facebook groups, Instagram stories).

EHR messaging

After filtering for study eligibility, we sent electronic study announcements through the EPIC patient portal to inform patients about the study. We also distributed study flyers and brochures in cancer center clinics. Both the study announcements sent through EPIC as well as all flyers and brochures had the link or quick response (QR) code to the study Research Electronic Data Capture (REDCap) screening survey for interested patients to learn more about the study and confirm eligibility.

Dermatology practice network

A partnership was formed with QualDerm Partners, a national network of dermatologists and skincare providers. QualDerm staff supported both passive and active recruitment approaches. Passive strategies, which required minimal research staff effort, included making flyers available in clinic waiting rooms and displaying electronic messages on digital boards. Active approaches involved mass email messaging to select providers and patients across the network. An institutional review board (IRB)-approved recruitment email was sent network-wide to providers, encouraging them to inform eligible patients about the study. QualDerm also implemented a direct patient messaging system which automatically emailed potentially eligible patients study details and a link to the REDCap eligibility screener. Although these messages were automated, they are considered active because staff time was needed to draft appropriate messaging, program them into the system, and define the target populations. This resulted in greater staff burden than the passive strategies.

Screening and enrollment

An online REDCap eligibility screener was used to prescreen potential participants. This screener was distributed through all recruitment channels via a QR code and URL. The survey automatically ended if a potential participant selected a response that made them ineligible at any point. All submitted screeners were carefully reviewed by research staff to identify and exclude bots or fraudulent entries. Several data quality safeguards were put in place to prevent fraudulent entries. These safeguards included limitations on the total number of REDCap entries allowed at a given time, which was monitored closely by research staff. In addition, Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA), a type of challenge-response test used to determine whether the user is a human or a bot, was also implemented to ensure that respondents were legitimate and not automated (13). Potential participants who met eligibility criteria through the online screener were contacted via phone by the research staff for additional verification. Once contacted by the study team, prospective participants either declined, were deemed ineligible or, if eligible, were sent an online consent and digital survey. After completion of the survey, participants were considered enrolled.

Statistical analysis

The analysis included 230 participants, a subset of approximately 300 individuals recruited for the parent RCT, which was the sample size calculated a priori to have power of at least 95% to detect intervention effects on the primary behavioral outcomes of the mySmartSkin intervention based on expected longitudinal retention rates. Enrollment yields and participant characteristics across non-cancer registry recruitment sources were summarized using descriptive statistics (counts and percentages). Enrollment rates were calculated as the proportion of individuals who completed the eligibility screening and subsequently enrolled in the study. Participant demographics were stratified by recruitment source, with between-source comparisons conducted using analysis of variance (ANOVA) for continuous variables and Chi-squared tests for categorical variables. Because the number of potential participants reached differed across recruitment sources and were often unknown (e.g., social media views, number of flyers seen by patients), total reach is not available to include.

Ethical statement

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Rutgers University Institutional Review Board (Study ID: Pro2022000948) and informed consent was obtained from all individual participants.

Results

Eligibility survey responses

Our REDCap screening survey yielded a total of 600 completed and valid responses (Figure 1). Among the 600 completed screeners, 348 (58%) screened eligible. Of the 252 individuals (42.0%) determined to be ineligible, common reasons included having received a melanoma diagnosis more than five years prior to screening, being diagnosed with stage IV disease, or currently undergoing treatment for an active melanoma diagnosis. After excluding these ineligible respondents, 348 individuals who met all study eligibility criteria were contacted for eligibility confirmation and consenting.

Figure 1 CONSORT of participants from non-registry sources. SSE, skin self-examination.

Study enrollment

Of the 348 screening eligible participants, 33 individuals (9.5%) actively declined to participate in the study. Another 21 individuals (6.0%) were identified as ineligible upon further screening, and 4 individuals (1.1%) did not complete the consent form and baseline survey. Furthermore, 25 individuals (7.2%) are currently in the enrollment process. Overall, a total of 230 participants (66.1%) consented, completed the baseline survey, and enrolled in the study from the three non-registry sources. In addition, over 1,700 entries from the screener pool have been removed to date due to poor quality, including incomplete, duplicate, or fraudulent responses. Because fraudulent entries were removed after screening and before personal contact the recruitment source for these entries could not be definitively determined. Based on the timing of batches of screeners received relative to postings, it was apparent that the majority of fraudulent entries originated from Facebook recruitment efforts.

Enrollment by recruitment source

Two hundred and thirty participants were enrolled via non-registry sources (Figure 2). This group included 122 participants (53.0%) referred by dermatology practice network outreach. Additionally, 56 participants (24.3%) were recruited through social media platforms (e.g., Facebook, Instagram). Another 46 participants (20.0%) were identified and enrolled via outreach efforts utilizing EHR messaging through a cancer center. Six participants (2.6%) entered the study through other sources (e.g., clinicaltrials.gov, friend referrals). Because they were not enrolled through our three primary sources, they were excluded from analysis.

Figure 2 Participants enrolled by non-registry source.

Participant demographics

Participant characteristics differed significantly by non-registry recruitment source (Table 1). Participants recruited via social media were significantly younger than those recruited through dermatology practice networks or EHR messaging (P<0.001) and were more likely to be women (P<0.001). Employment status varied by recruitment source (P<0.001), with social media recruitment yielding more full-time employed participants and fewer retirees than other sources. Insurance type also differed (P=0.001), with a higher proportion of privately insured participants recruited via social media. Cancer stage at diagnosis varied across sources (P=0.006), with social media recruitment yielding a greater proportion of participants diagnosed at stage I or higher. No significant differences were observed for race, ethnicity, minority status, education, income, or marital status.

Discussion

The recruitment outcomes of the mySmartSkin Study demonstrate the feasibility of leveraging three primary non-cancer registry recruitment sources. Each source had distinct strengths and limitations. Outreach through a dermatology practice network resulted in the greatest enrollment. These efforts, which included direct patient messaging, provider-targeted email campaigns, and the integration of printed materials into clinical settings, may have successfully leveraged clinical trust into research participation. Capitalizing on the provider-patient relationship offers a strategic opportunity to enhance engagement through an existing foundation of trust. Moreover, partnering with a large, well-established healthcare network likely increased credibility and participant comfort, further supporting enrollment. Additional advantages of this recruitment source included no direct study costs, minimal staff burden, and a low incidence of fraudulent entries. Although collaborating with a large dermatology practice network offered substantial benefits, this approach also introduced challenges. Finding a willing partner, developing the partnership, and the involvement of multiple departments within the network (e.g., marketing, medical records, clinical offices, providers, senior administration) required several layers of approval prior to recruitment initiation. Additionally, reliance on a private dermatology practice network may have contributed to the lack of racial diversity in our sample, as access to dermatologic care is known to be limited in general and particularly among some racial and ethnic minority groups (14). This highlights a potential trade-off of leveraging high-trust clinical networks: while they may improve enrollment and data quality, they may also limit representativeness and generalizability across more diverse populations.

Social media outreach, which resulted in the recruitment of nearly a quarter of the non-cancer registry cohort, also proved valuable. One key strength was that melanoma-based influencers and organizations provided trusted information and resources to their followers/members, which allowed us to tap directly into our target population in a credible way. These influencers and organizations also supported recruitment at no cost to the study, making it a highly cost-effective strategy.

Despite extensive reach and click-through rates, the paid Facebook ad campaigns yielded limited enrollment, as seen with other studies (15). Platform policy changes, particularly Meta’s restrictions on advertisement targeting, reduced the efficiency of the Facebook ads by limiting reach to our desired population. For social media strategies, fraudulent entries were a significant concern requiring both the implementation of safeguards to protect data integrity and substantial staff time to manually review and remove inauthentic responses. These findings are consistent with the work of Lei and colleagues who found that suspicious entries accounted for up to 82.69% of total survey entries (16). While social media recruitment is often considered a low-cost strategy, the staff time required for data verification and cleaning suggests that hidden costs may offset initial ad spending. Investigators should carefully weigh these trade-offs and consider whether staff have the bandwidth to manage and verify data before relying heavily on social media strategies, which are known to produce a high number of fraudulent entries. However, the literature describes automated safeguards that could be used, and artificial intelligence presents a novel option to help prevent and identify fraudulent responses going forward (17).

EHR recruitment contributed meaningfully to enrollment, reinforcing prior findings that electronic health data can be effectively leveraged to identify and engage eligible patients on a relatively large scale at no cost beyond staff time. However, the use of an electronic medical record portal posed its own challenges. Patient charts are not always up to date or accurate, which resulted in ineligible individuals receiving study information, creating inefficiencies by consuming staff time and potential patient confusion. Additionally, despite being delivered through a secure patient portal, the automated nature and absence of personal interaction may have reduced engagement or discouraged some individuals from learning about the study.

Eligibility screening data provides insight into the recruitment process. Of the 600 legitimate screeners that were received through the REDCap survey, about one-third were deemed ineligible based on study criteria. Given specific eligibility criteria, reaching a defined population may be challenging when utilizing broad recruitment sources. When using traditional methods, such as the cancer registries, contacts are initially filtered based on some eligibility criteria, approaching only those who seem to be eligible. Despite these exclusions, the screening and follow-up process successfully converted two-thirds of eligible individuals into enrolled participants, indicating a high yield from these recruitment efforts. Our conversion rates are consistent with other related studies (18). A scoping review focusing on recruitment and retention in randomized controlled trials of melanoma surveillance found that among 21 included trials an average of 75% (range, 24–100%) of those screened were eligible to participate and 63% (range, 24–95%) of those who were eligible went on to enroll (18).

Limitations of the parent study include recruiting convenience samples using strategies tailored to each recruitment source rather than consistent strategies across sources. However, findings regarding benefits and challenges are still informative, and optimal strategies likely require tailoring. As shown in Table 1, the use of non-cancer registry recruitment sources did not result in a demographically diverse sample, with an overwhelming majority of participants identifying as non-Hispanic white women. Melanoma is most common among non-Hispanic whites, volunteer influencers were more likely to be women, and women are more likely to engage with some types of social media (e.g., Instagram, Facebook) and enroll in research studies (19-21). This lack of racial, ethnic, and gender diversity may reflect limitations in the reach or accessibility of the recruitment methods used. Nonetheless, the sample did demonstrate some variation in education level, household income, and employment status, suggesting some socioeconomic diversity within the cohort. However, research continues to require improvement in efforts to recruit, engage, and retain diverse study participants.

The findings of this study align with literature supporting digital, health system, and provider-facilitated recruitment approaches in intervention research (4-6,22,23). Importantly, this study demonstrates that combining digital marketing with healthcare system-based engagement can yield large eligible samples of participants recruited for clinical trials.

Conclusions

The three recruitment sources utilized were helpful in reaching populations for cancer survivor intervention research. However, these approaches present unique challenges, such as the risk of potentially ineligible participants and fraudulent entries, which require verification safeguards to ensure the integrity of the recruitment process. Despite these challenges, the use of mostly unpaid sources and systems-based partially automated strategies demonstrated significant potential in the reach of interventions, offering participant enrollment that is aligned with real-world implementation. These findings underscore the value of leveraging such channels to engage a broad and geographically distributed audience.

Acknowledgments

Services, results, or products in support of the research project were generated by the Rutgers Cancer Institute Cancer Prevention and Outcomes Data Support Shared Resource. The authors acknowledge the following for their valuable contributions to this project: QualDerm Partners and John Albertini, MD (deceased), Carolina Lozada, MPH, Lisa Paddock, PhD, and Radiant Inc.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-25-61/rc

Data Sharing Statement: Available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-25-61/dss

Peer Review File: Available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-25-61/prf

Funding: This work was supported by the National Cancer Institute (No. R01CA265548).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://mhealth.amegroups.com/article/view/10.21037/mhealth-25-61/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Rutgers University Institutional Review Board and informed consent was obtained from all individual participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Jones EA, Shuman AG, Egleston BL, et al. Common Pitfalls of Head and Neck Research Using Cancer Registries. Otolaryngol Head Neck Surg 2019;161:245-50. [Crossref] [PubMed]
2. Kamarajah SK, Nathan H. Strengths and Limitations of Registries in Surgical Oncology Research. J Gastrointest Surg 2021;25:2989-96. [Crossref] [PubMed]
3. Mousavi R, Derakhshani N, Sheikhtaheri A, et al. Challenges and solutions of cancer registries: A scoping review. J Cancer Policy 2025;45:100607. [Crossref] [PubMed]
4. Hulbert-Williams NJ, Pendrous R, Hulbert-Williams L, et al. Recruiting cancer survivors into research studies using online methods: a secondary analysis from an international cancer survivorship cohort study. Ecancermedicalscience 2019;13:990. [Crossref] [PubMed]
5. Benedict C, Hahn AL, Diefenbach MA, et al. Recruitment via social media: advantages and potential biases. Digit Health 2019;5:2055207619867223. [Crossref] [PubMed]
6. Topolovec-Vranic J, Natarajan K. The Use of Social Media in Recruitment for Medical Research Studies: A Scoping Review. J Med Internet Res 2016;18:e286. [Crossref] [PubMed]
7. Ackermann DM, Dieng M, Medcalf E, et al. Assessing the Potential for Patient-led Surveillance After Treatment of Localized Melanoma (MEL-SELF): A Pilot Randomized Clinical Trial. JAMA Dermatol 2022;158:33-42. [Crossref] [PubMed]
8. Manne SL, Marchetti MA, Kashy DA, et al. mySmartCheck, a Digital Intervention to Promote Skin Self-examination Among Individuals Diagnosed With or at Risk for Melanoma: A Randomized Clinical Trial. Ann Behav Med 2022;56:791-803. [Crossref] [PubMed]
9. Marek AJ, Chu EY, Ming ME, et al. Piloting the Use of Smartphones, Reminders, and Accountability Partners to Promote Skin Self-Examinations in Patients with Total Body Photography: A Randomized Controlled Trial. Am J Clin Dermatol 2018;19:779-85. [Crossref] [PubMed]
10. Lamberti MJ, Dirks A, Kikuchi N, et al. An Examination of the Use of Patient Recruitment and Retention Tactics for Global Studies. MJH Life Sciences. 2024. Available online: https://www.appliedclinicaltrialsonline.com/view/an-examination-of-the-use-of-patient-recruitment-and-retention-tactics-for-global-studies
11. Song M, Elson J, Nguyen T, et al. Exploring trust dynamics in health information systems: the impact of patients’ health conditions on information source preferences. Front Public Health 2024;12:1478502. [Crossref] [PubMed]
12. Manne S, Heckman CJ, Frederick S, et al. A Digital Intervention to Improve Skin Self-Examination Among Survivors of Melanoma: Protocol for a Type-1 Hybrid Effectiveness-Implementation Randomized Trial. JMIR Res Protoc 2024;13:e52689. [Crossref] [PubMed]
13. Adesina AO, Ayobioloja PS, Obagbuwa IC, et al. An Improved Text-Based and Image-Based CAPTCHA Based on Solving and Response Time. Computers, Materials and Continua 2022;74:2661-75.
14. Narla S, Heath CR, Alexis A, et al. Racial disparities in dermatology. Arch Dermatol Res 2023;315:1215-23. [Crossref] [PubMed]
15. Greene A, Iloabuchi VC, Stoos E, et al. Social media as a recruitment tactic in melanoma education and protective behaviors in Latino patients. JEADV Clinical Practice 2024;3:844-9.
16. Lei F. Online recruitment for an online survey study: Our experience of dealing with fraudsters. Appl Nurs Res 2024;80:151854. [Crossref] [PubMed]
17. Mayer C, Tryon R, Ricks S, et al. Preventing fraudulent research participation: Methodological strategies and ethical impacts. J Genet Couns 2025;34:e70048. [Crossref] [PubMed]
18. Ackermann DM, Bracken K, Hersch JK, et al. Participant recruitment and retention in randomised controlled trials of melanoma surveillance: A scoping review. Contemp Clin Trials Commun 2025;44:101461. [Crossref] [PubMed]
19. American Cancer Society. Key Statistics for Melanoma Skin Cancer. 2025. Available online: https://www.cancer.org/cancer/types/melanoma-skin-cancer/about/key-statistics.html
20. Saphner T, Marek A, Homa JK, et al. Clinical trial participation assessed by age, sex, race, ethnicity, and socioeconomic status. Contemp Clin Trials 2021;103:106315. [Crossref] [PubMed]
21. Pew Research Center. Social Media Fact Sheet: Which social media platforms are most common? Pew Research Center. 2024. Available online: https://www.pewresearch.org/internet/fact-sheet/social-media/
22. Wang RR, Schweitzer JB, Hernandez S, et al. Strategies for recruitment and retention of adolescent and young adult cancer patients in research studies. J Clin Transl Sci 2023;7:e240. [Crossref] [PubMed]
23. Whitaker C, Stevelink S, Fear N. The Use of Facebook in Recruiting Participants for Health Research Purposes: A Systematic Review. J Med Internet Res 2017;19:e290. [Crossref] [PubMed]